A.I. was center stage at the 2024 Nobel Prizes, with Demis Hassabis sharing the Chemistry prize and Geoff Hinton sharing the Physics prize. Chem and Physics seems weird for A.I. though, no? Today's episode explains.
THE PRIZES
Widely considered the most prestigious award in their respective field, Nobel Prizes are administered by the Nobel Foundation in Sweden in the following fields:
Physics
Chemistry
Physiology or Medicine
Literature
Peace
Economics
The first five have been awarded since 1901, thanks to the will of dynamite inventor Alfred Nobel. The sixth, in Economics, was established in 1969 by Sweden's central bank.
Regardless of the field, Nobel laureates are those who have "conferred the greatest benefit to humankind." And, in the past year, the Nobel Foundation has noticed that A.I. has been making a big positive impact on humankind (there are of course negatives and risks too, as Hinton himself has been pointing out a lot lately).
Skimming the list of six prizes above, there is no Nobel Prize directly related to computing (the most prestigious award in computer science is the Turing Award, which was jointly awarded to Hinton alongside fellow Deep Learning pioneers LeCun and Bengio in 2018). But, that didn't stop the Nobel Prize committee from finding a way to recognize the achievements of those behind world-transforming artificial intelligence!
PHYSICS
The justification for awarding Hinton (a cognitive and computer scientist who's now Professor Emeritus at the University of Toronto) with a Nobel Prize in Physics yesterday is provided by Hinton sharing it with John Hopfield, an American physicist and emeritus professor at Princeton. A portion of Hopfield's work is classified as "biophysics", including his work on the development of artificial neural networks (ANNs), which are inspired by the physiology (and I guess the biophysics??) of biological brain cells. Hinton was pivotal in devising techniques for making ANNs "deeper" (that is, having more layers of processing), earning them the moniker "deep neural networks" (i.e., that carry out "deep learning") and Hinton the moniker of "godfather of A.I." Today, deep learning underlies essentially all of the cutting-edge A.I. capabilities, including those near-magical "reasoning" capabilities of state-of-the-art Large Language Models like OpenAI's o1.
CHEMISTRY
The justification for awarding *Sir* Demis Hassabis today is more straightforward. The Chemistry prizes are frequently awarded for biological (well, at least biochemical) breakthroughs and Hassabis (with co-laureate John Jumper) led development of AlphaFold, a Google DeepMind A.I. system that far exceeds human capacities at predicting protein structure, with tremendous biological research and healthcare implications.
The SuperDataScience podcast is available on all major podcasting platforms, YouTube, and at SuperDataScience.com.